Back-merge v2.1.0-rc.1 from main into develop - #5247
Merged
Conversation
…5115) * Bump ReportGenerator from 5.5.5 to 5.5.6 --- updated-dependencies: - dependency-name: ReportGenerator dependency-version: 5.5.6 dependency-type: direct:production update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com> * Update Examples link in README.md (#5104) * Reframe layout docs around responsive UI and instructional style (#5110) * Initial plan * docs: reframe layout docs around responsive UI Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/a6b913bd-38e5-4d80-ac23-9dc8e6716045 Co-authored-by: tig <585482+tig@users.noreply.github.com> * docs: polish responsive layout wording Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/a6b913bd-38e5-4d80-ac23-9dc8e6716045 Co-authored-by: tig <585482+tig@users.noreply.github.com> * docs: tighten instructional sentence style Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/d453216b-6d43-4261-aa23-0dad98805e91 Co-authored-by: tig <585482+tig@users.noreply.github.com> * docs: fix remaining instructional prose Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/d7be6123-896e-4b95-8e3f-0cee1d6d35d1 Co-authored-by: tig <585482+tig@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: tig <585482+tig@users.noreply.github.com> * Fixes #5114. Remove main from publish branch triggers; use tags only Main branch pushes were triggering the publish workflow without a release tag, causing GitVersion to produce invalid NuGet versions (e.g. 2.0.0-5420). - Remove main from branches trigger; stable releases now publish only on v* tag push - Develop branch pushes continue publishing pre-release packages automatically - Update template dispatch conditions to match tag-based flow Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Copilot <198982749+Copilot@users.noreply.github.com> Co-authored-by: tig <585482+tig@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
* Add tests for ShadowStyle inheritance and reset behavior Added two unit tests to ShadowTests.cs: - ShadowStyle_Getter_Does_Not_Inherit_From_SuperView ensures a child view does not inherit ShadowStyle from its SuperView and that round-tripping the property does not create a MarginView or add shadow thickness. - Setting_ShadowStyle_Null_Resets_ShadowSize verifies that setting ShadowStyle to null resets the MarginView's ShadowSize to Size.Empty, preventing residual state. * Fixes #5088 - Changed Margin.ShadowStyle to only use the local value and reset ShadowSize to Size.Empty when unset. Adjusted related tests to match new behavior. Refactored ShadowStyleDemo window and editor initialization for clarity and consistency, updated shadow window positioning and scheme usage, and improved child view arrangement. * Update Tests/UnitTestsParallelizable/ViewBase/Adornment/ShadowTests.cs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
…ut (#5083) * Fixes #5072. TableView reserves space for later columns when laying out A wide column (e.g. a long Description) previously consumed all available viewport width and pushed subsequent columns off-screen. Each visible column now reserves at least its header width (or configured MinWidth) during layout so wide columns are clamped instead of starving later columns. The reservation only inspects column headers and styles — it never iterates cell rows — so paginated or very large ITableSource implementations remain performant. * Tighten TableView column-bounds test to use strict inequality Right-edge check now uses < Viewport.Width instead of <= per code review. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * Restore foreach loop in TableView.CalculateContentSize Use a manual index counter rather than converting the loop to for, and revert the unrelated borderWidth local-variable extraction. The fix only needs the index to compute the slice of remaining columns to reserve space for. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * Updated table col width calculation and added wide example * Fixed tests and updated comment wording based on copilot pr feedback --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Co-authored-by: Tig <tig@users.noreply.github.com>
…iteration unnecessarily. (#5123)
… custom ColorGetter (#5128) * Fixes #5075. TableView ShowVerticalCellLines=false renders incorrectly with custom ColorGetter When ShowVerticalCellLines is false and a ColumnStyle.ColorGetter is in use (e.g. FileDialog), the 1-char gap between cells retained the row scheme's attribute, producing visible "stripes" through custom-colored cells. TruncateOrPad intentionally renders cells to (Width - 1) chars, leaving 1 char for the cell boundary. RenderRow now extends the cell's own background into that gap when the separator is the default invisible space. RenderSeparator skips drawing in the same case (RenderRow owns the gap). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * FileDialog: use ShowVerticalCellLines=false now that #5075 fix is in. Per PR review on #5128, switch FileDialog's TableView to render without vertical cell lines so its custom ColorGetter (file-type colors) shows through cleanly. The TableView fix in this PR ensures the 1-char gap between cells now picks up the cell's own background instead of the row scheme's, so removing the lines no longer leaves color stripes. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> Co-authored-by: BDisp <bd.bdisp@gmail.com>
* Revert "Fixes #5132. TableView Home/End navigation with FullRowSelect (#5133)" This reverts commit 6361f33. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Docs: Document Home/End rebinding for FullRowSelect in TableView Adds a tip to the FullRowSelect section showing how to rebind Home/End to Command.Start/Command.End for row-level navigation, referencing the UICatalogRunnable pattern. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
#5131) * Adds concurrency test for OutputBufferImpl race condition Adds AddStr_And_ClearContents_Concurrent_DoesNotThrow test that reproduces the race condition in OutputBufferImpl where ClearContents replaces the Contents reference while AddStr is concurrently iterating. See #5130. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Adds concurrency tests, benchmark, and plan for #5130 - Expanded OutputBufferImplConcurrencyTests with 4 tests covering: AddStr+ClearContents, FillRect+ClearContents, Move+AddStr+ClearContents, and a three-way concurrent test - Added OutputBufferBenchmark for perf baseline (AddStr, FillRect, ClearContents, SetSize, TypicalDrawCycle) - Added implementation plan in plans/ Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Refactor OutputBufferImpl for thread safety Replaced locking on Contents with a dedicated _contentsLock object to ensure safe concurrent access to buffer state (Contents, DirtyLines, Clip, etc.). Refactored buffer management methods (ClearContents, SetSize, FillRect, Move, IsValidLocation) to use the new lock. Updated FillRect and AddGrapheme to avoid race conditions. Improved documentation and comments for concurrency. Moved SetCellUrl/GetCellUrl implementations under the lock. In OutputBufferImplConcurrencyTests, reordered and moved helper methods for clarity; no test logic changes. No public API changes, but internal concurrency is now robust. * Refactor StringExtensions for style and consistency Refactored methods to use expression-bodied members and improved code style (constants, variable declarations, and formatting). Added and then removed `IsSurrogatePair` and `MakePrintable` extension methods, resulting in no net functional changes. All updates are stylistic and organizational. * Fix race condition in FillRect(Rectangle, char) overload The char overload called Move() + AddRune() without holding _contentsLock, allowing concurrent threads to corrupt Col/Row between the two calls and write characters to wrong positions. Fix: delegate to FillRect(Rectangle, Rune) which holds the lock atomically. Added test proving the race (detected wrong-position writes in <13ms before fix). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Added null check in OutputBufferImpl.cs WriteGrapheme. Introduced locking in RuneExtensions.cs for thread-safe Unicode width calculation. --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: BDisp <bd.bdisp@gmail.com>
…bleSelection.SelectedCell` (#5142) * Initial plan * Rename TableSelection.Cursor to TableSelection.SelectedCell Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/8d9d1a5e-97bb-4c6e-862a-948119c94fca Co-authored-by: tig <585482+tig@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: tig <585482+tig@users.noreply.github.com>
…inal character in v2.0.1 (#5146)
* Initial plan * Add custom header styling for TableView (Options B+C) - Add HeaderColorGetterArgs class for contextual header color info - Add HeaderColorGetterDelegate for per-column header color override - Add ColumnStyle.HeaderColorGetter property - Add TableStyle.HeaderScheme property for global header scheme - Update RenderHeaderMidline to apply header styling - Update CharacterMap scenario to italicize sorted column header - Add unit tests for both per-column and global header styling Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/157d9f83-14ee-4f74-9d4a-a6b879105d50 Co-authored-by: tig <585482+tig@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: tig <585482+tig@users.noreply.github.com> Co-authored-by: Tig <tig@users.noreply.github.com>
* Fix NativeAOT configuration cloning and warning cleanup Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix config converter regressions and serialization cleanup Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Remove stale AOT propagation annotations Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix NativeAOT configuration cloning and warning cleanup Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix config converter regressions and serialization cleanup Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Remove stale AOT propagation annotations Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix PR 5134 config review feedback Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Simplify AttributeJsonConverter style parsing Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Correctly serialize null properties in ScopeJsonConverter --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Co-authored-by: Tig <tig@users.noreply.github.com>
* Initial plan * Improve DropDownList keyboard: Space to open, Up/Down to navigate, CollectionNavigation support Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/c07c6615-2cdd-40a1-b0c0-86f0f31c26bb Co-authored-by: tig <585482+tig@users.noreply.github.com> * Address code review: split compound conditions into separate guard clauses Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/c07c6615-2cdd-40a1-b0c0-86f0f31c26bb Co-authored-by: tig <585482+tig@users.noreply.github.com> * Code cleanup --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: tig <585482+tig@users.noreply.github.com> Co-authored-by: Tig <tig@users.noreply.github.com>
…eadOnlyList<string>?> (#5160)
…#5150) * Initial plan * Fixes #5103. Disable NerdFonts default in FileDialogExamples scenario Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/49264ca7-5e02-4e48-aa49-e81b08ddcf23 Co-authored-by: tig <585482+tig@users.noreply.github.com> --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: tig <585482+tig@users.noreply.github.com> Co-authored-by: Tig <tig@users.noreply.github.com>
…ity (#5170) (#5179) Key.Equals(object?) included Handled in the comparison while GetHashCode returned only _keyCode.GetHashCode(), violating the .NET contract that equal objects must have equal hash codes. This broke KeyBindings (a ConcurrentDictionary<Key, KeyBinding>) when Handled flipped between storing and lookup. Handled is per-event state, not part of binding identity, so it is removed from the equality comparison. KeyEqualityComparer already used KeyCode-only equality and is consistent. Two existing tests in KeyTests.cs that asserted the buggy behavior (Equals_ShouldReturnTrue_WhenEqual and Equals_Handled_Changed_ShouldReturnFalse_WhenNotEqual) have been updated to assert the correct contract-honoring behavior. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv Co-authored-by: Claude <noreply@anthropic.com>
…eptable (#5189) CLAUDE.md and .claude/rules/testing-patterns.md disagreed on the AI marker convention (one said "// Claude - Opus 4.5", the other "// CoPilot - ChatGPT v4"). Empirically both are used: Tests/UnitTestsParallelizable + Tests/UnitTests.NonParallelizable contain ~105 files with "// Claude" markers and ~93 with "// CoPilot". Update both guidance files to list the two forms as equally acceptable and tell reviewers not to flag inconsistency between them. Triggered by Copilot review nits across PRs #5177-#5188. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv Co-authored-by: Claude <noreply@anthropic.com>
…) (#5186) * fix: UnixRawModeHelper safe restore on crash + valid-state guard (#5164) When the input thread crashes before the driver Dispose() runs, the Unix terminal is left in raw mode and the user's shell is left unusable. This change adds three safety nets so the original termios is restored on any abnormal exit path: - Finalizer: last-resort restore if Dispose() is bypassed (e.g. fatal native crash, immediate process abort). - AppDomain.ProcessExit hook registered after a successful TryEnable; unregistered in Dispose() to avoid leaking the helper for the lifetime of the process. - Console.CancelKeyPress hook so Ctrl+C also restores cleanly. Also adds a _haveSavedTermios guard so Restore() is a hard no-op unless tcgetattr previously succeeded - preventing an uninitialized termios struct from ever being written back to the terminal. Tests in UnitTests.Parallelizable cover: - Restore() without TryEnable() is a no-op and does not throw. - Restore() is idempotent. - Dispose() without TryEnable() is a no-op and double-Dispose is safe. - The type has a finalizer (last-resort restore contract). - TryEnable() failure paths leave _haveSavedTermios = false. The full crash-during-input-thread scenario is integration-only (requires a real tty); the unit tests cover the observable contract. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv * fix: check tcsetattr return; document CancelKeyPress limitation (#5164) Address Copilot review feedback on PR #5186: - Restore() and the finalizer now check the tcsetattr return code. On non-zero return, log the errno via the existing Logging infrastructure. Restore() leaves IsRawModeEnabled set to true when the syscall fails so callers and the finalizer know the terminal was not restored. The finalizer logs only and never throws. - Document the Console.CancelKeyPress limitation: with ISIG disabled in raw mode, a keyboard Ctrl+C arrives as a 0x03 byte rather than SIGINT, so the hook is unlikely to fire while raw mode is active. The subscription remains as belt-and-braces defense for hosts that do not disable ISIG and for externally-delivered SIGINT events; the primary safety nets are AppDomain.ProcessExit and the finalizer. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv * fix: drop unreachable finalizer from UnixRawModeHelper (#5164 review) The finalizer in UnixRawModeHelper was dead code: the instance is rooted by its AppDomain.CurrentDomain.ProcessExit and Console.CancelKeyPress event-handler subscriptions, so it can never become finalization-eligible while those hooks are wired. Drop the finalizer and the now-unneeded GC.SuppressFinalize call. Restoration relies on explicit Dispose as the primary path with the ProcessExit hook as the safety net for normal process exit; CancelKeyPress remains a best-effort fallback. Update the class-level remarks and the obsolete reflection-based finalizer test accordingly (6 tests -> 5). https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv --------- Co-authored-by: Claude <noreply@anthropic.com>
…) (#5188) * docs: clarify SessionBegun/SessionEnded as token lifecycle hooks (#5162) https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv * docs: tighten SessionBegun/SessionEnded XML accuracy (#5162) Address Copilot review feedback on PR #5188 by aligning the SessionBegun and SessionEnded XML docs with the actual ordering and contract in ApplicationImpl.Begin/End. - SessionBegun summary: describe the real firing point (after the token is pushed onto SessionStack and TopRunnable is set, but before IsRunningChanged/IsModalChanged fire) instead of "immediately after token construction". - SessionEnded summary: drop "token-disposal hook"/"about to be released" framing; SessionToken is not IDisposable. Describe the real state (state has unwound, IsRunningChanged(false) has fired, Runnable is null) without overstating the SessionStack invariant (End only pops when wasModal && popped == token). - Replace "manually dispose of the SessionToken" guidance in both remarks blocks with the actual caller responsibility: subscribe to SessionEnded and call End(SessionToken) themselves. - End(SessionToken) remarks: replace "disposes the sessionToken" with "clears SessionToken.Runnable on sessionToken" to match the code. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv --------- Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: BDisp <bd.bdisp@gmail.com>
…ose (#5163) (#5185) * fix: ApplicationImpl.Invoke throws NotInitializedException after Dispose (#5163) Both `Invoke(Action)` and `Invoke(Action<IApplication>)` overloads silently dropped actions when called before `Init()` or after `Dispose()`: the thread-affinity fast path is skipped (no main thread or top runnable) and the `TimedEvents.Add` fallback queues onto a stopped queue. Add an `Initialized` guard at the top of both overloads so the post- Dispose / pre-Init case fails loudly with `NotInitializedException`, matching the pattern already used by `Run(IRunnable, ...)` in the same file. The deeper thread-affinity race between the fast-path read of `TopRunnableView` / `MainThreadId` and concurrent shutdown is intentionally out of scope - that requires a redesign noted in the issue and is not covered here. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv * fix: address review comments on Invoke disposed-throws (#5163) - Pass nameof(Invoke) to NotInitializedException so the formatted "{member} cannot be accessed before Initialization" message is grammatically correct. - Add ArgumentNullException.ThrowIfNull(action) to the non-generic Invoke(Action) overload so a null action fails deterministically (consistent with Begin/Run argument validation). - Wrap Init in try/finally in the AfterDispose tests so a thrown Init does not leak the process-wide SynchronizationContext. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv --------- Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: BDisp <bd.bdisp@gmail.com>
…#5181) * fix: ConcurrentDictionaryJsonConverter rejects duplicate keys (#5173) The Read implementation was calling TryAdd and ignoring its return value, so duplicate keys in input JSON were silently dropped. Now throws a JsonException on duplicate, mirroring the DictionaryJsonConverter fix. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv * style: normalize space-before-paren in ConcurrentDictionaryJsonConverter (#5173 review) Apply project's space-before-paren convention to the Read method body, matching DictionaryJsonConverter<T>. Addresses review feedback on PR #5181 where the new TryAdd duplicate-key check used the convention but adjacent calls in the same method body did not. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv --------- Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: BDisp <bd.bdisp@gmail.com>
Per maintainer feedback, DrawOuterBoundary was a debug helper added during Region's development phase, not a supported public API. There are zero internal callers across the framework (verified via grep); the only consumer was the RegionScenario UICatalog demo, which now iterates GetRectangles() and adds rectangle edges to the LineCanvas directly. Shared edges between adjacent rectangles overlap; LineCanvas line-style joining handles the visual result, matching how Border/Adornment already draw. The 30+ structural smoke tests in DrawOuterBoundaryTests.cs are removed; none used DriverAssert or verified visual output. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: BDisp <bd.bdisp@gmail.com>
…5177) The KeyBindings constructor passed a binding factory to CommandBindingsBase that dropped the key parameter, calling `new KeyBinding(commands, source)` where `source` was bound to the View? overload's target/data slot. Result: every binding created via Add(key, ...) had Key = null and Target = null. Forward the key (and explicit nulls for target/data) to the (commands, key, source, target, data) constructor so KeyBinding.Key is populated. https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv Co-authored-by: Claude <noreply@anthropic.com>
The XOR branch in Region.Combine mutated `this` via Exclude(region) and then computed (region - this) against the already-mutated `this`, producing a non-symmetric result. Snapshot both operands first and compute (thisOriginal \ regionOriginal) ∪ (regionOriginal \ thisOriginal). https://claude.ai/code/session_01CdoS4YQ7v9aNPZaNsYnxjv Co-authored-by: Claude <noreply@anthropic.com>
# Conflicts: # GitVersion.yml # Terminal.Gui/Configuration/ConfigurationManager.cs # Terminal.Gui/Configuration/DeepCloner.cs Co-authored-by: tig <585482+tig@users.noreply.github.com>
Agent-Logs-Url: https://github.com/gui-cs/Terminal.Gui/sessions/743a69dd-85a4-4023-9d1f-003b2729230c Co-authored-by: tig <585482+tig@users.noreply.github.com>
Release v2.1.0-rc.1
tig
added a commit
that referenced
this pull request
May 7, 2026
The fix from #5243 ("Fixes #5259. AOT: use source-generated JSON serializer for dictionary construction in DeepCloner") was reverted when release/v2.1.0-rc.1 merged origin/main back at 94dca98. That merge had conflicts in DeepCloner.cs and ConfigurationManager.cs and the resolution picked main's pre-#5243 versions, dropping: - 27 lines from DeepCloner.cs (the SourceGenerationContext.GetTypeInfo fallback in CreateDictionaryInstance, plus its leading comment) and re-adding 54 lines to ConfigurationManager.cs: - CloneHardCodedPropertyValue and the per-type clone helpers (CloneKeyBindings, ClonePlatformKeyBinding, CloneKeyArray) that #5243 had removed in favor of generic DeepCloner handling - the call site change from DeepClone to CloneHardCodedPropertyValue PRs #5246 (rc.1 → main) and #5247 (rc.1 backmerge → develop) carried the broken state forward. v2.1.0-beta.1 was tagged before the merge (no fix); v2.1.0-rc.1 was tagged after but doesn't have it either. Repro: `dotnet publish Examples/NativeAot -c Release -r win-x64 --self-contained -p:PublishAot=true` against current develop crashes at module init with `MissingMethodException: No parameterless constructor defined for type Dictionary<ColorName16, string>`. Same crash a downstream AOT consumer (clet) hit and tracked as #5239 → #5240 → #5242 → #5243 → #5248. This commit checks DeepCloner.cs and ConfigurationManager.cs out of 11d6e27 (the original PR #5243 merge), restoring the source-gen fallback and removing the per-type helpers. Verified locally: - `dotnet build Terminal.Gui/Terminal.Gui.csproj -c Release` → 0 errors - clet AOT publish + `clet.exe --version` runs cleanly (was crashing against rc.1; now exits 0 with the local TG nupkg pinned) Refs #5239, #5242, #5243, #5248, #5259. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced May 8, 2026
This was referenced May 16, 2026
This was referenced May 25, 2026
This was referenced Jun 15, 2026
This was referenced Jun 29, 2026
This was referenced Jul 13, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Automatic back-merge after release v2.1.0-rc.1. Merge this to keep
developin sync withmain.